300 Journal of Medicinal Chemcstp, 1975, Vol 18, No. ,3
Horcharrlt.
(13) (a) C. Y . Chiou, J. P. Long, J. G. Cannon, and P. D. Armstrong, J . Pharmacol. Erp. Ther., 166, 243 (1969): (b) W. L. Nelson and R. Wilson, J . Med. Chem.. 14, 169 (1971); (c) N. J. Lewis, K. K. Barker, R. M. Fox, and M . P. Mertes, ibid., 16, 156 (1973); (d) J. B. Robinson. B. Belleau, and B. Cox, ibid., 12, 848 (1969); (e) B. Belleau and P . Pauling, ibid., 13, 737 (1970); (f) J. B. Kay, J. B. Robinson, B. Cos, and D. Polkonjak, J. Pharm. Pharmacol., 22,214 (1970). (14) E. Shefter and D. L. Tripgle, .Vaturfz (London), 227, 1354 (1970). (15) R. W. Baker, C. H. Chothia, P . Pauling, and T. .J. Petcher. Nature (London), 230,439 (1971). (16) G. H. Cocolas, E. C. Robinson. and LV L. Dewev. J .Vl~lecl Chem , 13,299 (1970)
Wu
(17) R. J . Radna, D. L. Beveridge, and A. L. Bender, J . Arncr. Chem. Soc., 95,3831 (1973). (18) D . W . Genson and R. E. Christoffersen, .J. Ampr.. (’hem Soc., 95,362 (19731, 119) J. L. Hogg, Department of Chemistry. T h e liniversitv of Kansas, unpublished results. (20) R. E. Bowman and H. H, Stroud, .I. Chcm. .Sot,, l:M2 ( 19j0 I . ( 2 1 ) A. H. Beckett, 9.J. Harper, and .J, W. Clitherow. J . I’hnrm. E)hurmacol.. 15, 349 (1963). (221 G. H . Cocolas, E. C . Robinson. and LV. 1,. Lkwev. .I .\fd Chem., 13,29911970). (23) R. T. Major and H. T. Bonnett. J . A r n w Chem .’lo(,.. 37, 2125 (1935).
Pot ential Inhibitors of S-Adenosylmethionine- Dependent Methy1transferases. 3. Modifications of the Sugar Portion of S-Adenosylhomocysteine Ronald T. Borchardt*,+and Yih Shiong Wu Department of Biochemistry, McCollum Laboratories, Uniuersit>
of
Kansas, Labrence, Kansas 66044 Receiued October 2, 1974
Structural analogs of S-adenosyl-L-homocysteine (L-SAH), with modification in the ribose portion of the molecule, have been synthesized and their abilities to inhibit catechol 0-methyltransferase (COMT), phenylethanolamine Nmethyltransferase (PNMT), histamine N-methyltransferase (HMT), and hydroxyindole O-methyltransferase (HIOMT) have been investigated. From these studies it was concluded that, in general, the 2’-hydroxyl and 3’-hydroxyl groups of the ribose moiety of SAH play crucial roles in the binding of this molecule to most methyltransferases. However, several interesting exceptions to this strict structural specificity have been observed. While S-3’-deoxyadenosyl-L-homocysteine produced no inhibition of H M T and HIOMT, it produced strong inhibition of the and S-5’-[9-(aratransmethylation catalyzed by PNMT and COMT. Likewise, S-2’-deoxyadenosyl-~.-homocysteine binofuranosyl)adenyl]-L-homocysteine had little or no effect of COMT, HMT, and HIOMT but were potent inhibitors of PNMT. The significance of these data relative to the nature of the SAH binding sites and the potential for in L I L ’ O differential inhibition of methyltransferases will be discussed.
S-Adenosyl-L-homocysteine (SAH)t produces strong product inhibition of most S-adenosylmethionine (SAM)dependent methy1transferases.l In the preceding papers of this we described the synthesis and enzymatic evaluation of SAH analogs with modifications in the homocysteine or base portions of the SAH molecule. From the results of these studies2J it was concluded that there exist at least four functional groups on SAH, which play a primary role in its binding to methyltransferases. These points of attachment appear to be the terminal carboxyl, the terminal amino and the sulfur atom of the homocysteine portion, and the 6-amino group of the adenine moiety. In an effort to further elucidate the nature of the intermolecular forces involved in enzymatic binding of SAH, we have synthesized a series of SAH derivatives with minor modifications in the sugar portion of this molecule (Chart I). Coward and coworkers4,spreviously prepared analogs in which the ribose moiety of SAH was replaced by a cyclopentyl group, a 2,3-dihydroxycyclopentyl group, or a fivecarbon acyclic bridge. However, these analogs were nearly inactive as inhibitors of SAM-dependent methyltransferases. In the present study we have made very minor changes in the ribose moiety of SAH in an effort to elucidate the important structural features of this portion of SAH in its binding to methyltransferases. Using these sugar-modified Whis work was done during the tenure of an Established Investigatorship 01’ the American Heart Association.
‘Abbreviations used are SAM, S-adenosyl-t.-methlonine; SAH, S-aderiosvl-r.-h[,mocysteine; 2’-deoxy-SAH, S-2’-deoxyadenosyl-l,-homocysteine: :{‘-deoxy-SAH,S-3’-deoxyadenosyl-L-homocysteine; SAraAH, S-6’-[9-(arahinnfuranosyl)adenyl]-L-homocysteine; COMT, cat,echol 0-methyltransfer ase (E.C. 2.1.1.6); PNMT, phenylethanolamine N-methyltransferase (E.C. 2.1.1); HMT. histamine h’-methyltransferase (E.C. 2.1.1.8); HIOMT. hydroxyindole 0-methyltransferase (E.C. 2.1.1.4); K,s, inhibition constant lor the slope.
Chart 1. Sugar-Modified Analogs of SAH Synthesized to Probe the Binding Sites on COMT, PMNT, HMT, and HIOMT
Ri
R2 R
SAH
on
OH
”3 E
2 ’ -deoxy-SA11
OH
Ii
H
3 ‘ -deoxy-SAH
ii
OH
H
OH
H
OH
Compd
SAraAH
1
2
SAH analogs (Chart I) as probes of the active sites of COMT, PNMT, HMT, and HIOMT, we have delineated the contribution of the ribose moiety in the enzymatic binding of SAH, which is the subject of this paper.
Experimental Section Melting points were obtained on a calibrated Thomas-Hoover IJni-Melt and were corrected. Microanalyses were conducted on an F & M Model 185 C. H, N analyzer, The University of Kansas, Lawrence, Kan. Unless otherwise stated, the ir, nmr, and uv data were consistent with the assigned structures. Ir data were recorded on a Beckman IR-33 spectrophotometer, nmr data on a Varian Associates Model T-60 spectrophotometer (TMS), and uv data on a Cary Model 14 spectrophotometer. Scintillation counting was done on a Beckman LS-150 scintillation counter. Tlc were run on Analtech silica gel GF (250 w ) or Avicel F (250 f i ) . Spots were detected hy visual examination under U V light and/or ninhydrin for compounds containing amino moieties.
S-Adenosylmethionine-Dependent Methyltransferases Materials. SAM-14CH3 (New England Nuclear, 55.0 mCi/ mmol) was diluted to a concentration of 10 fiCi/ml and stored a t -20'F. SAM iodide (Sigma) was stored as a 0.01 M aqueous stock solution. Phosphate buffers were prepared as 0.5 M stock solutions. The following compounds were commercially available from the indicated sources: 2'-deoxyadenosine, 3'-deoxyadenosine, and S benzyl-L-homocysteine (Sigma). 1'-8-D-Arabinofuranosyladenine was a gift from Dr. H. Wood, Drug Research and Development, Division of Cancer Treatment, NCI. S2'-Deoxyadenosyl-L-homocysteine (Z'-Deoxy-SAH). 5'O-p-Toluenesulfonyl-2'-deoxyadenine was prepared from 2'-deoxyadenosine and p-toluenesulfonyl chloride according to the procedure previously described by Robins, et aL6l7 The 5'-tosylate (435 mg, 1.07 mmol) was condensed with S- benzyl-L-homocysteine (216 mg, 96 mmol) in Na and liquid NH3 according to the previously described p r o c e d ~ r e .The ~ * ~crude ~ ~ 2'-deoxy-SAH was purified by thick-layer chromatography on cellulose (Analtech, 1000 f i ) eluting with EtOH-Hz0 (3:l). The product was removed from the cellulose by extraction with HzO followed by lyophilization. The lyophilized product was homogeneous on tlc [cellulose, EtOH-HzO (3: 2)] and was crystallized (HzO-acetone) to yield 240 mg (68%): mp 191'; nmr (Dz0) 6 7.93, 7.77 (2 s, 2 H, CZ-H, Cs-H), 6.00 (t, 1 H, Clc-H), 3.80 (m, 1 H, C3,-H), 4.20 (m, 1 H, C4,-H), 3.48 (t, 1 H, C,H), 2.10-2.77 (m, 6 H, Cz-Hz, Cs,-Hz, and C,-Hz), and 1.33-2.10 (m, 2 H, CpHz). A n d . ( C ~ ~ H Z O O ~ HzO) N S S C, H, N. S3'-Deoxyadenosyl-L-homocysteine (3'-Deoxy-SAH). 3'Deoxyadenosine (50 mg, 0.18 mmol) was dissolved in 4 ml of dry pyridine, freshly distilled from barium oxide, and the resulting solution was cooled to ea. 0-5' in an ice-salt bath. Freshly recrystallized p-toluenesulfonyl chloride (55 mg, 0.28 mmol) was added in portions over a 1-hr period, after which the solution was kept for 8 hr a t 5-10' and another 4 hr a t ambient temperature. The reaction mixture was poured into an ice-water-CHCl3 mixture and the CHC13 layer was separated and washed with cold 1.0 N HzS04, cold saturated NaHC03, and ice-HzO. The CHCb solution was dried (Na2S04), filtered, and concentrated under reduced pressure. The concentrated CHC13 solution was added dropwise to cold hexane and the white solid which formed was filtered and recrystallized (CHC13) to yield 30 mg (41%) of the desired 5'-O-p-toluenesulfonyl-3'-deoxyadenine: mp 158-160'; nmr (DMSO-de) 6 8.10 (s, 2 H, CZ-H,Cs-H), 7.69 (d, 2 H, phenyl), 7.27 (d, 2 H, phenyl), 5.87 (s, 1 H, C1,-H), 4.60 (br, 1 H, Cz-OH), 4.53 (m, 2 H, C2,H, C4,-H), 4.03-4.37 (m, 2 H, C5,-H), 2.31 (s, 3 H, -CH3), and 1.92-2.30 (m, 2 H, C3'-Hz). Spectral data were consistent with the proposed structure, but an accurate analysis could not be obtained because of the instability of the compound when recrystallization was attempted. 5'-O-p-Toluenesulfonyl-3'-deoxyadenine(30 mg, 0.074 mmol) was condensed with S-benzyl-L-homocysteine (25 mg, 0.148 mmol) in Na and liquid NH3 as described above for 2'-deoxy-SAH. The crude 3'-deoxy-SAH was purified by thick-layer chromatography on cellulose (Analtech, 1000 p ) , double developing with EtOHH20 (3:l). The product was removed from the cellulose by extraction with Hz0 followed by lyophilization to yield 10 mg (37%). The product was homogeneous on tlc [cellulose, EtOH-H20 (3:2)] and was crystallized [water-acetone) to yield 3'-deoxy-SAH: mp 211' dec; nmr (D20) 6 7.93, 7.97 (2 s, 2 H, C2-H, Cs-H), 5.78 (d, 1 H , CI-H), 4.23 (m, 2 H, C p H , Cds-H), 3.43 (t, 1 H , C,-H), 2.47-2.78 (m, 2 H, Cy-Hz), 2.13-2.47 (m, 2 H, C,-Hz), and 1.50-2.10 (m, 4 H, C3'-HZ, Cp-Hz). A d . (C14HZ004N&.HzO) C, H, N. 9-( 5'- 0-pToluenesulfony1-@-D-arabinofuranosy1)adenine (1). Compound 1 was prepared by a modification of the procedure of Robins, et d eAnhydrous 1-8-D-arabinofuranosyladenine(1.27 g, 5 mmol) was dissolved in 94 ml of boiling pyridine, freshly distilled from barium oxide. To the pyridine solution, cooled to ca. 0-5' in an ice-salt bath, was added dropwise over 1hr a solution of p-toluenesulfonyl chloride (1.44 g, 7.5 mmol) in 36 ml of CHC13pyridine (1:l). After addition was completed the solution was stirred a t 0' for 8 hr. The reaction mixture was poured into a solution of saturated aqueous NaHC03 a t 0-4' and the CHC13 layer was separated and the aqueous phase extracted with two 20-ml portions of cold CHC13. The combined CHC13 fractions were washed with water and dried (NazS04). After filtration the CHC13 was removed under pressure a t 0' and the residue was dissolved in 1 ml of CHC13 and added dropwise to cold hexane. The precipitate was collected by filtration to yield 1.04 g (51%): mp 156-157'; nmr (DMSO-d~)6 8.17, 7.98, (2 S, 2 H, Cz-H, Cs-H), 7.80 (d, 2 H, phenyl), 7.40 (d, 2 H, phenyl), 6.28 (d, 1 H, CIS-H),5.57-6.03 (m, 2 H , Czz-OH, Cy-OH), 4.36 (d, 2 H, C5,-H), 3.80-4.50 (m, 3 H, (22,-H
Journal of Medicinal Chemistry, 1975, Vol. 18, No. 3 301 C3,-H, Cg-H), and 2.38 (s, 3 H, -CH3). Various attempts to crystallize this material for C, H, and N analysis resulted in decomposition. Therefore, it was converted to the diacetyl derivative 2. 9-( 2',3'- O-Diacetyl-5'- 0-ptoluenesulfonyl-8-D-arabinofuranosy1)adenine (2). To a solution of 5'-tosylate 1 (500 mg, 1.2 mmol) in 4 ml of pyridine was added 4 ml of AczO and the solution was stirred a t 0' for 6 hr. The reaction mixture was poured into an ice-HZO-CHCls mixture and the CHC13 layer was separated and the aqueous phase extracted several times with cold CHC13. The CHC13 solutions were combined, washed with H20, and dried (Na2S04). The CHC13 was removed under reduced pressure and the desired tosylate 2 crystallized (hexane-CHCl3) to yield 430 mg (71%):mp 226-228'. Anal. ( C Z ~ H Z ~ N ~ O S )N. C,SH, Attempted Synthesis of S5'-[9-(Arabinofuranosyl)adenyl]L-homocysteine (SAraAH). (a) From 9-(5'- O-pToluenesulfonyl-8-D-arabinofuranosy1)adenine (1). The tosylate 1 (95 mg, 0.23 mmol) was condensed with S-benzyl-L-homocysteine (53 mg, 0.23 mmol) in Na and liquid NH3 according to the previously described p r o c e d ~ r e After . ~ ~ ~the ~ ~NH3 had evaporated, the residue was dissolved in 1 ml of HzO, neutralized (1.0 N HCl), and cooled (ea. 0-4'). The solid which formed was filtered and recrystallized (Hz0) to yield 46 mg (80%), mp 298' dec. The ir, uv, and nmr spectra indicated that the product obtained was 9-(2',5'-anhydro-/3-D-arabinofuranosy1)adenine(3) (lit.6*9mp was reported as darkened near 220' and liquified with decomposition a t 300'). Anal. (C10HllN503) C, H, N. (€3) From 9-(2',3'-O-Diacety~-5'-O-p-toluenesulfonyl-,9-Darabinofuranosy1)adenine (2). Using the standard condensation p r o c e d ~ r e tosylate , ~ ~ ~ ~2~(218 mg, 0.41 mmol) was allowed to react with S-benzyl-L-homocysteine (90 mg, 0.4 mmol) in Na and liquid "3. After the NH3 had evaporated, the residue was dissolved in 5 ml of cold HzO and the aqueous layer extracted several times with CHC13. The aqueous layer was lyophilized and the product purified by thick-layer chromatography on cellulose (Analtech, 1000 p ) eluting with EtOH-H20 (3:l). Fraction A (Rf 0.64) was extracted with HzO and crystallized (HzO) to yield 65 mg (66%) of 9-(2',5'anhydro-P-D-arabinofuranosy1)adenine(3), mp 298'. Fraction B (Rf 0.23) was extracted from the cellulose with HzO and the HzO removed by lyophilization to yield 45 mg of the crude product, which contained the desired SAraAH and homocysteine. This product was further purified by thick-layer chromatography on cellulose (Analtech, 1000 p ) eluting with a mixed solvent system containing nine parts of EtOH-AcOH-H20 (50:3:5) and one part of phosphate buffer, pH 7.0 (0.02 M),and then rechromatography on cellulose eluting with EtOH-Hz0 (3:l). The desired SAraAH was extracted with H20 and the H20 removed by lyophilization to yield 9 mg (6%) of the product: mp 235' dec; nmr (CF3COzD) 6 9.68, 9.35 (2 s, 2 H, CZ-H,Cs-H), 7.10 (s, 1 H, Clr-H), 5.50-6.07 (m, 3 H, CZJ-H,Cy-H, C4,-H),5.17-5.50 (m, 1 H, C,-H), 4.05-5.02 (m, 4 H, Cy-Hz, C,-Hz), and 3.13-3.87 (m, 2 H, CpH2). (The unexpected peak positions observed for SAraAH appear to be the result of the nmr solvent (CF~COZD), since shifts to similar positions were observed for SAH and homocysteine when the spectrums were recorded in CF3COzD); uv X max (HgO) 260 nm ( e 10,200). Anal. (Ci4H2005NeS) C, H, N. Enzyme Purification Assay. The enzymes used in this study were purified from the following sources according previously described procedures: COMT,l0J1 rat liver (male, Sprague-Dawley, 180-200 9); PNMT,12 bovine adrenal medulla (Pel-Freez Biologicals); HMT,13 guinea pig brain (Pel-Freez Biologicals); and HIOMT,I4 bovine pineal glands (Pel-Freez Biologicals). COMT, PNMT, HMT, and HIOMT were assayed and the analogs of SAH evaluated as inhibitors using the radiochemical techniques described in the preceding papers of this series.2~~ The assay mixtures for each of the methyltransferases contained SAM-14CH3 (0.05 pCi) and SAM (1.0 mM). The acceptor molecules and their final concentrations in the assay mixtures were as follows: COMT, dihydroxybenzoate (2.0 mM); PNMT, DL-P-phenylethanolamine (1.0 mM); HMT, histamine (1.0 mM); and HIOMT, N-acetylserotonin (1.0 mM). Processing of the kinetic data was accomplished as previously d e s ~ r i b e d . ~ ~ ~ ~ ' ~ . ' ~ ~ ~ ~
Results and Discussion Chemistry. The various sugar-modified analogs of SAH prepared i n this s t u d y a r e listed in C h a r t I. 2'-Deoxy-SAH a n d 3'-deoxy-SAH were p r e p a r e d b y t h e general synthetic procedures outlined in a preceding p a p e r of this series.3 The appropriate nucleosides were converted to their 5'-to-
302
Jottrnal ofMedicinal Ghemistry, 1975, Vol. 18, No. 3
Scheme I CHvqyadenine
R=H
Borchardt, Wu
Table I. Inhibition of COMT, PNMT, HMT, and HIOMT by Sugar-Modified Analogs of SAH”
t
InhibItor;
Na-NH, S-benzyl- homocysteine
I OH 3
C o mpd
SAH
I
OR /R=H 2, R = -C(=O)CH, R =a=O)CH, +
Na-NH,, S-benzyl- L -homocysteine
3
+ SAraAH
(
inhibition
COIICI~,
m.11
2’-Deoq-SAH
0.2 2.0 0.2
3’-Deoxy-SAEI
2.0 0.2
SAraAH
2.0 0.2 2.0
COMTPNMT HMT HIOMT
39 87 0 2 22 59 0 0
49 92 9 45
52 90
11 33
40 89
71
7
0 5 2 0 6 14
25 10 29 5
94
13 sylates and condensed directly with S- benzyl-L-homocys“COMT, PNMT, HMT, and HIOMT were purified and assayed teine to yield the desired SAH derivatives (2’-deoxy-SAH as described in the Experimental Section except in each case the and 3‘-deoxy-SAH). During these condensation reactions it SAM concentration = 1.0 mM. bThe inhibitors were prepared in was not necessary to protect the 2’- or 3’-hydroxyl groups aqueous stock solutions (10.0 ~ m o l / m l ) . of the nucleoside 5’-tosylates. However, difficulty was encountered in the synthesis of SAraAH, because the condensation of 9-(5’-~-p-to~uenesulfonyl-fi-D-arabinofuranoSAH, and SAraAH, thus showing little specificity for the sy1)adenine (1)with S-benzyl-L-homocysteine in sodium and 2’- and 3’-hydroxyl groups of the sugar moiety. This apliquid NHs afforded 9-(2‘,5’-anhydro-fi-~-arabinofurano-pears to be particularly true for the 3’-hydroxyl group, sy1)adenine (3) rather than the desired SAraAH (Scheme I). since 3’-deoxy-SAH is as potent as SAH itself in inhibiting In fact, this intramolecular cyclization of tosylate 1 to this enzyme. However, the 2’-hydroxyl group may be some2’,5’-anhydronucleoside 3 took place in sodium and liquid what important in binding, since 2’-deoxy-SAH and SAraNH3 alone. This intramolecular cyclization had previously AH were less potent inhibitors than SAH. With COMT been observed to occur in ethanol with sodium methoxide the 2’-hydroxyl group appears necessary for the optimal at room t e r n p e r a t ~ r e . ~In , ’ ~an attempt to prevent this cybinding of SAH, since a’-deoxy-SAH and SAraAH were clization, 5’-tosylate 1 was converted to 9-(2’,3’-O- diacetylcompletely devoid of inhibitory activity. However, the 3’5’-O-p-toluenesulfonyl-~-~-arabinofuranosyl)adenine (2) hydroxyl group must not be directly involved in the bindusing acetic anhydride in pyridine. Condensation of 2 with ing, since 3‘-deoxy-SAH exhibited substantial inhibitory S - benzyl-L-homocysteine afforded a mixture of the 2’,5’activity toward this enzyme, anhydronucleoside 3 and the desired SAraAH. That the The kinetic patterns of inhibition of COMT, PNMT, and major product was again the 2‘,5‘-anhydronucleoside 3 HMT by SAH, 2’-deoxy-SAH, 3’-deoxy-SAH, and SAraAH suggests that the acetyl-protecting groups of 2 were rapidly were determined and the resulting inhibition constants are removed under the condensation conditions permitting listed in Table 11. In all cases linear competitive patterns of rapid cyclization to 3. However, displacement of the 5’-tosinhibition were observed when SAM was the variable subylate of 2 by L-homocysteine anion must have been comstrate. Of particular interest is the potent inhibitory activipetitive with the intramolecular cyclization, since a signifity of 3‘-deoxy-SAH toward PNMT, where the inhibition cant amount of SAraAH was also isolated. The 2’,5’-anhyconstant of 42.7 i 2.45 p M for this sugar analog is comparable to that for SAH (Kls = 29.0 f 2.84 p M ) . With 2’dronucleoside 3 appears to arise from an intramolecular displacement of the 5’-tosylate rather than from SAraAH, deoxy-SAH and SAraAH the inhibition constants toward since SAraAH was stable under the conditions used for PNMT are about ten times greater than that for SAH, inconsideration reaction. All of the SAH derivatives and their dicating that some of the interactions critical to formation of the PNMT-inhibitor complex have been lost by removal synthetic intermediates were characterized by their ir, nmr, of the 2’-hydroxyl or inversion a t C-2.3’-Deoxy-SAH is also and uv spectra, their chromatographic properties, and elea potent inhibitor of COMT, but nearly inactive with mental analyses. HMT. These apparent differences in the specificity for the In Vitro Inhibition Studies. The various sugar-modiribose portion of SAH a t the enzymatic binding sites could fied analogs of SAH, which were synthesized as part of this be potentially useful in obtaining differential inhibition in study, were tested as inhibitors of COMT, PNMT, HMT, uiuo and HIOMT and the results are shown in Table I. Included The competitive kinetic data discussed above for SAH for comparison are data for the inhibition of these enzymes by SAH. The enzyme showing the highest specificity for and 3’-deoxy-SAH inhibition of PNMT suggest that both the structural features of the ribose portion of SAH was of these inhibitors are binding to the same site on this enHIOMT, since none of the analogs listed in Table I inhibitzyme. To provide evidence in support of this assumption a study of the kinetics of multiple inhibition of PNMT by ed this knzyme appreciably. HIOMT appears to have an extremely high specificity for all of the structural features SAH and 3’-deoxy-SAH was conducted using the proceof SAH, since analogs with modifications in the amino acid dures of Yonetani and Theorall.ls As shown in Figure 1, a or base portions of this molecule also did not inhibit.2J series of parallel straight lines was obtained when reciproWith HMT the various sugar-modified SAH analogs cal velocities were plotted us. SAH concentrations at varyshowed very slight inhibitory activities, indicating that ing concentrations of 3‘-deoxy-SAH, indicating that these both the 2‘-hydroxyl and 3’-hydroxyl groups must be imtwo inhibitors are competing for the same site on PNMT. portant in binding to this enzyme. In contrast, PNMT, an Conclusions enzyme which exhibits high specificity for the structural The present study has attempted to elucidate the imporfeatures of the base and amino acid portions of SAH,2J was very susceptible to inhibition by 2’-deoxy-SAH, 3’-deoxytance of the 2’- and 3‘-hydroxyl groups of the ribose moiety
Journal of Medicinal Chemistry, 1975, Vol. 18, No. 3 303
S-Adenosylmethionine-DependentMethyltransferases
Table 11. Inhibition Constants for SAH, 2’-Deoxy-SAH, 3’-Deoxy-SAH, and SAraAH toward COMT, PNMT, and HMTa Inhibition Ki, i S.E.M.
COMT
PNMT
HMT
36.3 i 2.20
29.0 i 2.84 278 i 12.2 42.7 i 2.45 206 i 27
18.1 i 2.19
Inhibitor ~~
SAH 2 ’-Deoxy -SAH 3’-Deoxy-SAH SAraAH
pLM,
~~
138 f 31.2
2070 i 864
“COMT, PNMT, and HMT were purified and assayed as described in the Experimental Section. SAM concentration, 24-210 N M .bEach inhibitor showed linear competitive kinetics and the inhibition constants were calculated as previously described.2JJ1.15.16 CIf low inhibitory activity was observed from the preliminary studies, no attempt was made to determine the kinetic inhibition constants.
of SAH in the binding of this molecule to COMT, PNMT, HMT, and HIOMT. The SAH analogs of interest in this study were 2’-deoxy-SAH, 3’-deoxy-SAH, and SAraAH. From the inhibitory activities of these sugar-modified analogs, it can be concluded that HMT and HIOMT show strict specificity for the structural features of the ribose portion of SAH. However, the binding sites of PNMT and COMT appear to be less sensitive to changes in the ribose moiety, since strong inhibition of PNMT by 2‘-deoxy-SAH and 3’-deoxy-SAH and moderate inhibition of COMT by 3’-deoxy-SAH were observed. These interesting exceptions to the normal strict specificity for SAH binding may permit differential inhibition in uiuo of these catecholamine metabolizing enzymes from other methyltransferases. From the inhibitory activities observed for the various amino acid, base, and sugar-modified analogs of SAH prepared in our it can be generally concluded that methyltransferases show strict specificity for the structural features of SAH. However, several interesting exceptions to this strict specificity have been observed which may permit differential inhibition of methyltransferases in uiuo. Chart I1 shows a schematic representation of a possible enzymatic binding site for SAH. This proposed Chart 11. Proposed Binding Sites for SAH
h
/
/
-1;
P-------V A
o
lo
20 BAH].
o 30
40
o
IO‘
Figure 1. Reciprocal PNMT velocity us. SAH concentration with varying 3’-deoxy-SAH concentration. DL-P-Phenylethanolamine concentration, 1.0 mM. SAM concentration, 1.0 mM. Velocity = nmol of product/mg of protein/min.
the terminal carboxyl (site a), terminal amino (site b), and sulfur atom (site c) of the homocysteine portion; the 6amino group (site f) of the base portion; and the 3‘-hydroxyl (site d) and 2’-hydroxyl (site e) groups of the ribose portion of SAH. The binding of SAH through the terminal amino group (site b) and the 6-amino group of adenine (site f) appears to be particularly important for all methyltransferases studied. For the optimal binding of SAH to HMT and tRNA methyltransferase all of the functional groups shown in Chart I1 except the terminal carboxyl group (site a) and the configuration of the amino acid asymmetric carbon are required.2.20-22 For the binding of SAH to COMT, the sulfur atom (site c) and the 3’-hydroxyl group (site d) do not appear crucial for optimal binding; however, the other functional groups shown in Chart I1 are absolute requirernenk2 For PNMT the 2’-hydroxyl group (site e) and 3’-hydroxyl group (site d) of SAH do not appear to be required for maximum binding, since removal of either of these groups does not result in loss activity. However, all of the other attachment sites on SAH are crucial in optimal binding to PNMT. The enzyme HIOMT shows a very high specificity for all the structural features of SAH and only certain base-modified derivatives retain significant inhibitory a ~ t i v i t yThe . ~ differences in the specificity of the SAH binding sites, which have been detected in our studies, are being further explored in an attempt to design specific inhibitors of these enzymes. Acknowledgment. The authors gratefully acknowledge support of this project by a Research Grant from the National Institutes of Neurological Diseases and Stroke (NS10198) and by Contract HSM-42-73-8 from the National Institute of Mental Health. The excellent technical assistance of Bi-Shia Wu is gratefully acknowledged.
References
binding site for SAH is similar to that proposed by Zappia, et al.,19 for the enzymatic binding of SAM. The structural features of primary importance in the binding of SAH are
(1) R. T. Borchardt in “The Biochemistry of S-Adenosylmethionine,” E. Borek, Ed., Columbia Univeriity Press, New York, N.Y., in press. (2) R. T.Borchardt and Y. S. Wu,J . Med. Chem., 17,862 (1974). (3) R. T. Borchardt, J. A. Huber, and Y. S. Wu, J. Med. Chem., 17,868 (1974). (4) J. K.Coward and W. D. Sweet, J . Med. Chem., 15,381 (1972). (5) J. K.Coward and E. P. Slisz, J. Med. Chem., 16,460 (1973). (6) M.J. Robins, J. R. McCarthy, and R. K. Robins, Biochemistry, 5,224 (1966). (7) M. G.Stout, M. J. Robins, P. K. Olsen, and R. K. Robins, J. Med. Chem., 12,658 (1969).
304 Johrnal of Medicinal Chemistty, 1975, Vol. 18, No. 3
Fuller, Housh. Snoddy, Da). Mollo)
(8) W. Sakami, Biochern. Prep., 8,8 (1961).
(9) M. Hubert-Habart and L. Goodman, Can. J . Chem., 48, 1335 (1970). (10) R. Nikodejevic, S. Senoh, J. W. Daly, and C. R. Creveling, J . Pharmacal. Exp. Ther., 174,83 (1970). (11) R. T. Borchardt, J .
[email protected]., 16,377,383,387,581(1973). (12) R. J. Connett and N. Kirshner, J . Riol. Chem., 245, 329 (1970). (13) D. D. Brown, R. Tomchick, and J. Axelrod, J . Hiol. Chem., 234,2948 (1959). (14) R. L. Jackson and W. Lovenberg, J . Riol. Chem., 246, 2948 (1971). (15) W.U'. Cleland, Nature (London),198,463 (1963).
(16) G. N. Wilkinson, J . Biochem., 80,324 (1961). (17) M. G. Stout and R. K. Robins, J . Heterocycl. Chem., 8, 516 ( 1971). (18) T. Yonetani and H. Theorall, Arch. Riochem. Riophys., 106, 243 (1964). (19) V. Zappia, C. R. Zydek-Cwick, and F. Schlenk. J . Bioi. ('hem., 244,4499 (1969). (20) J. Hildesheim, ,J. F. Goguillon, and E. Lederer, FEBS Lett.. 30,177 (1973). (21) ,J. Hildesheim, R. Hildesheim, J. Yon, and E. Lederer, Hiochimie, 54,989 (1972). (22) J. Hildesheim, R. Hildesheim, P. Blanchard, G. Farrugia, and R. Michelot, Riochimie, 55,541 (1973).
Norepinephrine N- Methyltransferase Inhibition by Benzamidines, Phenylacetamidines, Benzylguanidines, and Ph.enylethylguanidines Ray W. Fuller,* Betty W. Roush, Harold D. Snoddy, William A. Day, and Bryan B. Molloy T h e Lilly Research Laboratories, Indianapolis, Indiana 46206. Received September 20, 1973 Norepinephrine N-methyltransferase (NMT) from rabbit adrenal glands was inhibited by benzylamine and phenethylamine analogs in which the nitrogen was replaced by an amidino or guanidino group. Mono and dichloro derivatives of benzamidines, phenylacetamidines, benzylguanidines, and phenethylguanidines were studied. The two most potent NMT inhibitors among the compounds examined were 2,3-dichlorobenzamidine and 3,4-dichlorophenylacetamidine, with pIso values of 5.55 and 5.36, respectively. These inhibitors were reversible and were competitive with norepinephrine as the variable substrate. They inhibited NMT from human, rat, and bovine adrenal glands but were slightly less effective against those enzymes than against the rabbit adrenal enzyme. In exercised rats, 2,3-dichlorobenzamidine had no significant effect on adrenal catecholamine levels. 3,4-Dichlorophenylacetamidineslightly reduced epinephrine levels in the adrenal glands of exercised rats, but the effect may have been due to release rather than inhibition of synthesis, since heart norepinephrine levels were also reduced significantly by that agent (which is from a chemical series known to release catecholamines). Thus, whereas these compounds are reasonably potent inhibitors of NMT i n uitro, they apparently are not effective in blocking enzyme activity in vivo.
Norepinephrine N - methyltransferaset (NMT) catalyzes the terminal step of epinephrine biosynthesis in the adrenal medulla. This enzyme is an ideal target for inhibiting epinephrine formation; an inhibitor of NMT would not affect the synthesis of dopamine or norepinephrine, which have physiologic functions of their own in addition to being precursors of epinephrine. The pharmacologic consequences of blocking NMT are unknown, since agents that specifically inhibit this enzyme in vivo have not been reported. For several years we have been searching for inhibitors of NMT and have found that phenethylamines and benzylamines are among the most effective inhibitors in ~ i t r o . " This ~ paper describes some compounds of those types in which the amine function has been replaced by an amidino or guanidino group. The compounds studied have the structures RCNH, or RNHCNH,
II
It
rlH 9H where R = phenyl, benzyl, or phenethyl. Inhibition by Analogs of Benzylamines. Table I shows the inhibition of NMT by chlorinated benzamidines. In-
."-
'This enzyme has previously heen referred to as phenylethanolamine methyltransferase.' However, norepinephrine is a much better substrate (lower K,) than phenyl ethanolamine*^ and is perhaps the only physiological substrate for the enzyme. Thus the 1972 recommendations of the International Union of Pure and Applied Chemistry and the International Union of Biochemistry list noradrenalin N- methyltransferase as their recommended name for this enzyme (E.C. 2.1.1.28).5 Following their recommendations and replacing norepinephrine (U. S . ) for noradrenalin (British), we intend to refer to this enzyme as norepinephrine N-methyltransferase (NMT for norepinephrine methyltransferase) in the future, despite the extensive use of the name phenylethanolamine N-methyltransferase in the prior literature.
Table 1. I n Vitro Inhibition (pI50 Values) of NMT by Benzylamines, a-Methylbenzylamines. and Benzamidines Aromatic substituent Series
None
2-C1
3-Cl
4-Ct
C,HjCH2NH, C6H,CHNH2
3 12 3 04'
4 66 5 29
5 07 4 72
4 07 3 70
4 97
6 23 6 42
CH, C,H,CNH,
2 62'
4 09'
4 25
3 02'
3 83
5 55
fl)?
(2)
13)
(4)
(5)
(6)
I1
"
3.4-C12,3-C1
a d isomer :3.17. i isomer 2.2. Marshallton Research Laboratories. i. Reference 8. Compound numbers are shown i n parentheses in this table and Table 11.
eluded for comparison are data for the inhibition by benzylamines and a-methylbenzylamines (taken from ref '7). The benzamidines were slightly weaker inhibitors of NMT than were the benzylamines, but the order of potency of the various chlorine-substituted isomers was generally similar. The 2,a-dichloro compound was the most active inhibitor in all three series. The 3-chloro compound was next most potent with the 2-chloro just behind it in the benzamidine series and the benzylamine series. Among the N methylbenzylamines, the 2-chloro compound was more potent than expected probably due to steric interaction between the a-methyl and the bulky 2-chloro substituent. The 4-chloro compound was least potent of the chloro compounds in all three groups.
